Professional Documents
Culture Documents
SLADE Paul G.†, LI Wang-pei, MAYO Stephen, SMITH R. Kirkland, TAYLOR Erik D.
(Eaton Electrical, Horseheads, NY 14845, USA)
†
E-mail: PaulGSlade@eaton.com
Received Dec.19, 2006; revision accepted Jan. 10, 2007
Abstract: The use of vacuum interrupters (VIs) as the current interruption component for switches, circuit breakers, reclosers
and contactors operating at distribution voltages has escalated since their introduction in the mid-1950’s. This electrical product
has developed a dominating position for switching and protecting distribution circuits. VIs are even being introduced into
switching products operating at transmission voltages. Among the reasons for the VI’s popularity are its compactness, its range of
application, its low cost, its superb electrical and mechanical life and its ease of application. Its major advantage is its
well-established reliability. In this paper we show how this reliability has been achieved by design, by mechanical life testing and
by electrical performance testing. We introduce the “sealed for life” concept for the VI’s integrity. We discuss this in terms of what
is meant by a practical leak rate for VIs with a life of over 30 years. We show that a simple high voltage withstand test is an easy
and effective method for monitoring the long-term vacuum integrity. Finally we evaluate the need for routine inspection of this
electrical product when it is used in adverse ambient environments.
Key words: Vacuum interrupter (VI), Reliability, Mechanical life, Electrical life
doi:10.1631/jzus.2007.A0335 Document code: A CLC number: TM5
80
the bellows. This voltage is generally below the
60 breakdown voltage of the open contact gap at at-
Vacuum
SF6
mospheric pressure; therefore it depends on the
40 Oil breakdown voltage dropping through the minimum of
Air the Paschen curve as air leaks into the VI. For an
20 excellent compilation of Paschen curve data, see
(Dakin et al., 1974).
0
1980 1990 2000 2010
VIs need to have no or very low chance of me-
Year chanical failure below their rated number of
close-open operations. Mechanical life tests were
Fig.1 Sales of different medium voltage circuit breaker performed on a set of VIs, and the distribution of the
technologies by year
life data was analyzed using three-parameter Weibull
distribution (Bruning et al., 1996) given by:
are required to mechanically open and close for a
large number of operations. A bellows made from β
x − x
thin stainless steel allows the contacts to open and f ( x) = 1 − exp 0
, (1)
close while retaining vacuum inside the interrupter. α − x0
The fatigue performance of the bellows limits the
mechanical life of the VI. The contact opening and where f(x) is the cumulative probability of failure, x is
closing operations stress the bellows, particularly the the mechanical life, α is the life where the fraction of
convolutions closest to the ends. In addition to the 1/e units have failed, β is the shape parameter of the
direct motion from the operation, the bellows oscil- distribution, and x0 is the threshold value below which
lates after the contact motion stops, further adding to the probability of failure is effectively zero. Fig.2
the wear on the bellows. The critical contact motion illustrates an example of this analysis. The plot shows
parameters that affect mechanical life are: the cumulative probability of failure as a function of
(1) steady state contact stroke or gap; the number of operations. This probability is the
(2) opening and closing speed; percentage that has failed at of before a particular
(3) motion damping at end of opening and mechanical life. The three-parameter Weibull distri-
closing stroke; bution reproduces the behavior of the data, and gives
(4) overshoot and rebound on opening; a threshold value of 43770, which is just slightly
(5) mounting resilience; below the lowest mechanical life measurement of
(6) contact bouncing on closing. 44096.
The mechanical life of Eaton Electrical VI’s was
evaluated by performing opening and closing opera- 99
Cumulative probability
50
mechanism that a VI may use in practice; e.g. that of a 25
switch, a circuit breaker or a contactor. The opening 10
and closing speeds reflect values recommended for 5
use in these mechanisms. For example, a VI that is to 2
be used in a vacuum circuit breaker will have its 1
4 5 6 7 8 9
contact gap set to the circuit breaker’s nominal gap, Mechanical life (×104)
plus the gap tolerance, plus the amount of overtravel.
The opening and closing speeds are adjusted towards Fig.2 Cumulative percentage of VIs experiencing me-
chanical failure as a function of the number of opera-
the maximum values. This gives the maximum de- tions. These results are for a particular bellows design
flection stress on the bellows. The testing pauses after and circuit breaker parameters. The threshold value is
every 100 operations, opens the contacts and applies a 43770 operations
Slade et al. / J Zhejiang Univ Sci A 2007 8(3):335-342 337
Most VIs are seldom operated. The electrical and forces the high current vacuum arc to move rapidly
mechanical wear the interrupters experience in ser- around the periphery of the contacts. The second,
vice is usually only a small fraction of their actual called the Axial Magnetic Field (AMF) contact,
capability. Table 1 gives an overview of the expected forces the high current columnar vacuum arc into a
life for a number of common VI applications. In high current diffuse mode.
general the VI will certainly perform its switching and When switching the diffuse vacuum arc that
protection function for most applications in excess of occurs when normal load currents (e.g. 630 A to 3150
the expected life of the equipment in which it resides. A) are interrupted, it is important to consider the de-
It is possible to develop bellows for mechanical re- posit of material eroded from the cathode and depos-
quirements beyond the VI life given in Table 1. For ited on the anode. Schulman et al.(1999) showed that
those applications where very frequent switching is when this is taken into account the effective erosion
needed, the number of operations should be moni- of the VI contacts in an ac circuit is considerably
tored and a regular maintenance schedule developed. reduced. They determined the effective erosion as a
The long-term reliability of the VI while switching function of the contact gap <g> divided by the contact
current has permitted these frequent switching ap- diameter <φ> as well as the effect of slots cut into the
plications to be implemented. contact’s surface. For example, a 62 mm diameter
Cu-Cr contact with <g>/<φ> and slots in 28% of the
Electrical life contact surfaces has an effective erosion of 13×10−7
Since the first introduction of the VI it has cm3/C. The life of this contact is given as a function of
demonstrated an exceptional electrical switching life. load current in Table 2. This contact life is much
The major advantage of the contact in vacuum is that larger than the 30000 operations expected from the
throughout its switching life the contact resistance normal operating life of a vacuum circuit breaker. VIs
remains nearly constant. There is no gas inside the VI developed for vacuum contactors have been designed
to change the chemical composition of the contact. with switching lives of up to 5 million electrical
Slade and Smith (2006) explored this life over the switching operations.
whole current range expected for VIs used in vacuum A 62 mm diameter TMF contact switching 25
circuit breakers. These VIs are designed to interrupt kA (RMS) has a switching life of 100 operations
the diffuse vacuum arc that results during load cur- (Slade and Smith, 2006). For currents >6 kA the life
rents switching (for currents ≤6 kA) and the high of this contact is inversely proportional to the square
current vacuum arc that results when switching fault of the current. The contact life as a function of current
currents (for currents >6 kA). In order to control the is shown in Fig.3. The AMF contact of the same
high current vacuum arc two contact structures were diameter has a much lower erosion rate for the same
successfully developed (Slade, 1984). The first, high current interruption. A comparison of the two
named the Transverse Magnetic Field (TMF) contact, contact designs is shown in Fig.4 for a 35 mm di-
Table 2 Life of a 62 mm TMF contact switching load Smith, 2006). It is thus possible to design VIs to have
currents until 3 mm of erosion is reached an outstanding operating life performance, which is
Circuit current Switching to reach greatly in excess of any former and any competing
(RMS) (kA) 3 mm contact erosion interrupter technology.
0.63 953835
1.25 480733 Vacuum integrity, sealed for life
In the early days of application of VIs, there was
1.60 375572
genuine concern about vacuum integrity. This might
2.00 300458 have been the result of the glass envelope used in the
2.50 240366 manufacture of these early VIs. Glass is vulnerable to
3.15 193532 permeation of both hydrogen and helium. At that time
some end users were told that the life expectancy of a
VI was 20 years. In the last few years we have heard
Inverse square calculation from 100 ops @ 25 kA
Diffuse vacuum arc calculations some utility customers replacing perfectly good VIs
Empirical measurement of erosion in their switchgear simply because it has been 20
106
years since the panel boards were installed. Our own
Life, number of operations
105 experiences in seeing our VIs used in the field all over
Mechanical life=30000 ops
Mechanical life=10000 ops the world, in vastly different environments, for well
104
over 40 years, showed that with proper vacuum
103 quarantine testing carried out in the VI manufacturing
plant, VIs are “sealed for life”. A VI can and should
102
continue to be used for as long as it passes a high
10 voltage AC withstanding test, showing that it still has
102 103 104 105 an acceptable vacuum. As far as we know, those VIs
Current (A)
we manufactured in 1968 are still operating reliably.
Fig.3 The contact life vs current (RMS) curve for 62 In fact, an analysis in 1998 of an Eaton (formerly
mm diameter TMF contacts inside a 102 mm diameter Westinghouse) VI that had been in the field since
vacuum interrupter 1970 showed an internal pressure of 1×10−3 Pa.
Once the VIs have gone through the final as-
4.0 sembly, evacuation and brazing process, the com-
Linear contact erosion (mm)
Leusenkamp, M., Hilderink, J., Lenstra, K., 1996. Field Cal- Slade, P.G., 1984. The vacuum interrupter contact. IEEE Trans.
culations on Epoxy Resin Insulated Vacuum Interrupters. CHMT, CHMT-7(1):25-32.
Proc. 17th Int. Symp. on Discharges and Electrical Insu- Slade, P.G., 1997. Vacuum interrupters: the new technology
lation in Vacuum, Berkeley, USA, p.1065-1069. for switching and protecting distribution circuits. IEEE
[doi:10.1109/DEIV.1996.545528] Trans. on Industry Applications, 33(6):1501-1511.
Long, W., Smith, R.K., 2003. Are All Circuit Breakers Created [doi:10.1109/28.649962]
Equal? Not When It Comes to Some Generator Circuits. Slade, P.G., Smith, R.K., 2006. Electrical Switching Life of
Proc. Water Power Conference XIII, Buffalo, USA. Vacuum Circuit Breaker Interrupters. Proc. 52nd IEEE
Okawa, M., Tsutsumi, T., Alyoshi, T., 1987. Reliability and Holm Conf. on Electrical Contacts, Montreal, Canada,
field experience of vacuum interrupters. IEEE Trans. on IEEE Cat. No. 06CH37833, p.32-37. [doi:10.1109/HOLM.
Power Delivery, PWRD 2(3):799-804. 2006.284061]
Schulman, M., Slade, P., Loud, L., Li, W.P, 1999. Influence of VVPC (Varian Vacuum Products Catalog), 1996. Varian
contact geometry and current on effective erosion of Vacuum Products, Lexington, MA, p.201.
Cu-Cr, Ag-WC, and Ag-Cr, vacuum contact materials. Wilson, N., Beavis, L., 1976. Handbook of Vacuum Leak
IEEE Trans. Components and Packaging Technologies, Detection. Pub. American Vacuum Soc., p.55.
22(3):405-413. [doi:10.1109/6144.796543]
► Contribution requests
(1) Electronic manuscript should be sent to jzus@zju.edu.cn only. If you have any questions,
please feel free to visit our website (http://www.zju.edu.cn/jzus), and hit “For Authors”.
(2) English abstract should include Objective, Method, Result and Conclusion.
(3) Tables and figures could be used to prove your research result.
(4) Full text of the Science Letters should be in 3−4 pages. The length of articles and reviews is
not limited.
(5) Please visit our website (http://www.zju.edu.cn/jzus/pformat.htm) to see paper format.